The indirect value of flexibility across power generation, trading and procurement — and why it is worth measuring.
Flexibility business cases inside integrated utilities are typically built as a standalone P&L: balancing revenue, intraday arbitrage, an ancillary services line, a customer revenue share, less platform and onboarding cost. Those value pools are real, well understood, and increasingly competitive as ancillary prices compress and pure-play aggregators sharpen their terms.
For a utility that also owns generation, runs a trading book or serves a supply portfolio, there is a second set of value pools alongside them. In that setting flexibility also works as a constraint-cancellation layer underneath three other books, and what it does to the economics of those books is significant — indirect, harder to attribute, and still thinly covered in flexibility research.
They are worth counting. Here is what comes into view when the boundary is drawn a little wider.
The frame: flexibility as constraint cancellation
Flexibility is best understood not as a resource in its own right, but as a set of measures that relieve specific binding constraints in a system.[1] That framing transfers cleanly from a system to a portfolio.
A large integrated portfolio has known, expensive, structural rigidities: thermal units with ramp-rate limits, minimum-load floors and start-up costs; wind and solar with forecast error that is punished asymmetrically in imbalance settlement; reserve obligations that must be met by a plant that would rather be selling energy; risk limits and collateral requirements that bind hardest exactly when volatility is highest.
Distributed commercial and industrial flexibility — batteries, CHP, boilers, backup generation, shiftable industrial process load, behind-the-meter PV — has almost exactly the inverse technical signature. It is fast where the fleet is slow, granular where the fleet is lumpy, and geographically distributed where the fleet is concentrated.
A recent comprehensive review of virtual power plants and aggregators makes the same structural argument at system level: the value of aggregation is not additive, it comes from coordination across heterogeneous resources whose profiles are imperfectly correlated.[2]
Three value pools follow.
Value pool 1 — The generation book: capture price and the thermal core
The defensive bidding tax. Intermittent generation faces an asymmetric penalty structure: over-forecasting can lead to paying punitive imbalance, while under-forecasting may forfeit a high-price sale. Research quantified how renewable forecast errors propagate into imbalance volumes and spot price formation, and the asymmetry is real and material.[3] Desks respond rationally by bidding conservatively — commonly 85–90% of expected output rather than the P50. That discount is significant. It is a permanent, self-imposed haircut on realized capture price, paid every settlement period, to insure against a risk that could instead be covered physically.
A flexibility layer with credible, fast, contracted response changes the calculus: bidding closer to P50 is enabled because a physical correction is available that does not require buying back into a stressed intraday market. The revenue uplift shows up in the generation book, not in the flex P&L.
Protecting the thermal core. When large units are used for fine balancing and reserve provision, they operate away from their heat-rate optimum and absorb ramping and cycling duty. Technical work on thermal plant flexibility documents what that costs in fuel efficiency, component fatigue and maintenance-driven OpEx.[4] Displacing high-frequency modulation onto distributed assets lets thermal units sit at optimal set-points. Better fuel efficiency, less cycling damage, longer intervals between maintenance-driven outages.
Reserve opportunity cost. Another piece of research models the integrated management of energy and ancillary services in a pay-as-bid setting and shows the opportunity cost explicitly: capacity committed to reserve is capacity not selling energy at the moment it is most valuable.[5] Moving reserve compliance onto a distributed pool releases the merit-order assets back into the energy market.
Outage and maintenance elasticity. A distributed pool of contracted flexibility is short-term replacement capacity. It buys the option to schedule maintenance around market value rather than around a conservative calendar, and it covers short unplanned gaps without buying back at scarcity prices.
None of these four lines appear in a standalone flexibility business case. All four land in the generation book.
Value pool 2 — The trading book: convexity, intraday execution, and risk capital
This is where the indirect value is largest and least visible.
From hedge to real option. A financial hedge is symmetric and it costs premium. A physical flexibility layer is asymmetric: at each intraday decision point you choose whether to correct physically or transact in the market, based on which is cheaper at that moment. Downside is physically capped; upside from volatility is not. That is a convexity change in the P&L distribution, and convexity is worth more than the expected value of any single dispatch.
Intraday liquidity is thin exactly when you need it. Bergault and Cognéville’s work simulating sparse order books in intraday electricity markets is a useful corrective to the assumption that you can always trade out of a position.[6] Spreads widen and depth evaporates during precisely the events — outages, congestion, sharp forecast revisions — that create the imbalance you need to close. Internal physical dispatch bypasses the bid-ask spread, the market impact, the counterparty credit exposure and the clearing overhead entirely. Per correction the saving is small. Across thousands of corrections a year it compounds into a structurally lower cost of balancing.
Gate-closure behaviour. Participants without physical recourse flatten early and accept suboptimal prices. Participants with it can hold exposure closer to delivery, into the window where the largest price moves cluster. This is a behavioral change in the desk, enabled by a physical capability — and it is invisible in any model that treats the flexibility asset as a revenue line.
Risk capital and collateral. A review of VaR methodologies for energy portfolios is a reminder that risk limits are a binding commercial constraint, not just a compliance artifact.[7] A physical layer that demonstrably reduces imbalance tail exposure supports lower VaR add-ons and wider position limits — a larger deployable risk budget on the same capital. Dampened net exposure also reduces initial and variation margin requirements. That is a direct ROCE effect that shows up nowhere near the flexibility business case.
Firmness is probabilistic, not binary. The objection is always delivery certainty. The work on probabilistic flexibility aggregation for ancillary services provision addresses this directly: a pool of heterogeneous DERs is characterized by a distribution, not an on/off state.[8] Brattle’s Real Reliability analysis reached the same conclusion from the resource-adequacy side — a properly diversified virtual portfolio delivers reliability comparable to conventional capacity.[9] Aggregate firmness is an engineered property, and it is what qualifies the pool for premium, high-availability products.
Value pool 3 — The retail and C&I procurement book
The internal physical hedge. A supply book carries shape and volume risk that is conventionally hedged with financial instruments. C&I load you can actually move is a physical hedge against your own supply obligation — it reduces derivative reliance and, with it, margin and liquidity buffers.
Product upgrade. Flexibility converts “as-available” renewable output into products a corporate buyer will pay a premium for: firm delivery blocks, PPAs with stabilized shape and ramp guarantees. You are internalizing the cost of firming rather than passing volume risk to the offtaker or eating it in the merchant book.
Churn. This is the softest of the claims and should be treated as such. The logic is sound — technical integration into a customer’s operational assets raises switching cost, and a revenue-share relationship is structurally stickier than a price-per-MWh one — but the published evidence is thin. If you build a business case on retention uplift, insist on measuring it against a matched control group rather than asserting it.
Scale: the threshold below which none of this is true
The indirect value pools only open above a minimum effective scale, and this is where most pilots quietly fail. The test is not “how many MW have we signed” but “does the pool cover a meaningful fraction of the standard deviation of the renewable portfolio’s imbalance?” Below that threshold you have an interesting demonstration. Above it, the trading desk can change its bidding behavior — and the behavior change is where the value is.
The sizing exercise is straightforward and worth doing before anything else:
- Define precisely what you are hedging — renewable imbalance, reserve obligation, start-up and cycling cost, scarcity exposure.
- Quantify the stochastic variable and its distribution over the relevant horizons.
- Set the risk metric — VaR reduction, reliability target, reserve margin — and solve for the flexible capacity that materially shifts it.
Two structural notes. First, aggregate reliability is a function of the law of large numbers, so concentration is the enemy: a pool dominated by a handful of large sites behaves like a lumpy asset, not a portfolio. Second, the incumbent advantage here is real and underused. The trading desk, the dispatch systems, the risk framework and the market access are already built and already paid for. The marginal cost of the first MW of C&I flexibility for an integrated utility is far below that of a pure-play aggregator, which has to fund the entire J-curve. A survey of VPP operational strategies sets out how wide that required capability stack actually is.[10]
Market design is the constraint that remains. Minimum bid sizes, prequalification requirements and product definitions across ancillary services still exclude small-scale flexibility more than they need to — a barrier documented for the EU in the Utilities Policy review of small-scale flexibility in European markets.[11] Aggregation is partly a technical answer to a regulatory problem.
What to stress-test before believing any of this
Widening the boundary of a business case is exactly where analysis tends to go soft. Three things to be hard about:
Attribution. Indirect value is by construction difficult to attribute. If it cannot be measured, it cannot be defended in year two. Build the counterfactual before building the platform: a shadow portfolio, a documented pre-intervention bid-ratio baseline, an agreed method for booking imbalance cost avoided.
Governance. The value lands in books that did not pay for the asset. Generation, trading, and retail each capture part of it; the flexibility unit carries the cost. Most attempts at this die in the org chart long before they die in the economics. Internal transfer mechanisms are required.
The trust layer. Every claim above depends on dispatch and trading acting on telemetry from thousands of assets sitting behind customer firewalls. If that data is not verifiable, and if control commands are not authenticated to the device, then you have not built a portfolio hedge — you have built a correlated operational risk and pointed your trading desk at it. Device identity, data integrity and auditable measurement are load-bearing infrastructure here, not a procurement afterthought.
Conclusion
The direct revenue pools in flexibility are real, competitive and compressing. The indirect pools are structural, defensible, and sit almost entirely inside the boundary of what an integrated utility already owns. The firms that work this out first will not be the ones with the most MW under management. They will be the ones who counted correctly.
References
[1] The most-cited taxonomy of power system flexibility classifies it as a portfolio of measures — supply-side, network, storage and demand-side — each relieving a specific binding constraint on the integration of variable renewable generation, rather than as a single resource type. Lund, P.D., Lindgren, J., Mikkola, J., & Salpakari, J. (2015). Review of energy system flexibility measures to enable high levels of variable renewable electricity. Renewable and Sustainable Energy Reviews, 45, 785–807. https://doi.org/10.1016/j.rser.2015.01.057
[2] Kaiss, M., Wan, Y., Gebbran, D., Vila, C.U., & Dragičević, T. (2025). Review on Virtual Power Plants/Virtual Aggregators: Concepts, applications, prospects and operation strategies. Renewable and Sustainable Energy Reviews, 211, 115242. https://doi.org/10.1016/j.rser.2024.115242
[3] Goodarzi, S., Perera, H.N., & Bunn, D. (2019). The impact of renewable energy forecast errors on imbalance volumes and electricity spot prices. Energy Policy, 134, 110827. https://doi.org/10.1016/j.enpol.2019.06.035
[4] Agora Energiewende (2017). Flexibility in thermal power plants — with a focus on existing coal-fired power plants. https://www.agora-energiewende.org/publications/flexibility-in-thermal-power-plants
[5] Vannoni, A., & Sorce, A. (2024). Integrated energy and ancillary services optimized management and risk analysis within a pay-as-bid market. Applied Energy, 371, 123628. https://doi.org/10.1016/j.apenergy.2024.123628
[6] Bergault, P., & Cognéville, E. (2024). Simulating and analyzing a sparse order book: an application to intraday electricity markets. arXiv:2410.06839. https://arxiv.org/abs/2410.06839
[7] Halkos, G.E., & Tsirivis, A.S. (2019). Value-at-risk methodologies for effective energy portfolio risk management. Economic Analysis and Policy, 62, 197–212. https://doi.org/10.1016/j.eap.2019.03.002
[8] Jacobs, M., & Paolone, M. (2025). Probabilistic Flexibility Aggregation of DERs for Ancillary Services Provision. arXiv:2503.15378. https://arxiv.org/abs/2503.15378
[9] Hledik, R., & Peters, K. (2023). Real Reliability: The Value of Virtual Power, Volume I: Summary Report. The Brattle Group. https://www.brattle.com/wp-content/uploads/2023/04/Real-Reliability-The-Value-of-Virtual-Power_5.3.2023.pdf
[10] Roozbehani, M.M., Heydarian-Forushani, E., Hasanzadeh, S., & Elghali, S.B. (2022). Virtual Power Plant Operational Strategies: Models, Markets, Optimization, Challenges, and Opportunities. Sustainability, 14(19), 12486. https://doi.org/10.3390/su141912486
[11] Opportunities and challenges for small-scale flexibility in European electricity markets. (2023). Utilities Policy, 80, 101474. https://www.sciencedirect.com/science/article/abs/pii/S0957178722001412
Further context on the direct value pools: U.S. Department of Energy (2025), Pathways to Commercial Liftoff: Virtual Power Plants 2025 Update; RMI (2024), Power Shift: How Virtual Power Plants Unlock Cleaner, More Affordable Electricity Systems.